ABSTRACT While Salmonella VNP 20009 (VNP) has shown promise in tumor targeting in animal models it is becoming apparent that regrowth of tumors is the outcome since the bacteria preferentially grow in the necrotic core and fail to reach the vital periphery. In human trials of metastatic melanoma patients, even targeting itself was very poor and no tumor regression was observed. We aim to improve the tumor targeting and tumor regressive ability of VNP by conferring upon it the successful recognition potential of antibody mediated tumor targeting. We had previously attempted to do this using the lpp-OmpA surface display platform and poorly soluble scFv antibodies but have since found out that the combination caused bacterial growth arrest and restricted the therapeutic potential of engineered VNP. We hypothesize that VNP displaying small and highly soluble anti- CEA single domain antibodies via the ice nucleation protein (INP) will show enhanced targeting of CEA positive tumors without inhibiting bacterial invasion or replication and will therefore be able to regress tumors more effectively than parental VNP. Our specific aims are to; 1, Engineer VNP to display anti-CEA sdAb without impeding replication, yet capable of binding immobilized CEA in vitro; 2, Demonstrate engineered VNP is capable of being internalized in MCF-7 cells and is also capable of homogenous targeting in three- dimensional spheroidal cell culture models; 3, Demonstrate improved in vivo tumor targeting of engineered VNP in a mouse model of breast cancer, elucidating intra-tumoral distribution and potential for complete tumor regression. The multifunctional capabilities of VNP including mobility and ability to carry large payloads are well balanced by the ability to visualize single bioluminescent cells and facile elimination of the bacteria by antibiotics. As such, the potential for VNP as a highly controllable yet versatile cancer therapeutic is enormous so long as we can improve targeting and regressive capacity.